Organolithium/chiral Lewis base/BF3: A versatile combination for the enantioselective desymmetrization of meso-epoxides

Article abstract of DOI:10.1002/ejoc.200400745

BF₃ can be used in combination with organolithium/strong Lewis base complexes for the enantioselective nucleophilic ring-opening or the carbenoidic rearrangement of various meso-oxiranes with excellent yields and ee values of up to 87%. Mechanistic aspects of these reactions are considered. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2005.

Full text of DOI:10.1002/ejoc.200400745

FULL PAPER
Organolithium/Chiral Lewis Base/BF3: a Versatile Combination for the
Enantioselective Desymmetrization of meso-Epoxides
Emmanuel Vrancken,*[a] Alexandre Alexakis,[b] and Pierre Mangeney[a]
Keywords: Boranes / Carbenoids / Lithium / Lewis bases / Synthetic methods
BF3 can be used in combination with organolithium/strong
Lewis base complexes for the enantioselective nucleophilic
ring-opening or the carbenoidic rearrangement of various
meso-oxiranes with excellent yields and ee values of up to
87%. Mechanistic aspects of these reactions are considered.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
Since BF3 is known to promote the nucleophilic addition
of a wide range of organometallic reagents to epoxides in a
simple and effective way,[10–14] this approach appears to be
a practical and versatile methodology. Among all the chiral
ligands tested with organolithium reagents, strong Lewis
base donors such as amino alcohols, amino alkoxides, or
diamines give the best results in term of enantio-
selectivity.[15] However, as we started this work, the compat-
ibility between strong Lewis base/organolithium complexes
and strong Lewis acids had never been addressed.[16,17]
Therefore, we decided to explore the scope and the general-
ity of such a reaction. Here, we wish to report all our results
obtained during this study.[6]
Introduction
The concept of the combination of an organolithium rea-
gent and a chiral Lewis base for enantioselective reactions
has been widely explored. Similarly, the combination of an
organolithium and BF3 has also been extensively studied.[1]
Thus, we wondered if it would be possible to use a combina-
tion of these two methodologies (i.e. RLi + ligand + BF3) in
order to take advantage of the enantioselective possibilities
arising from the chiral ligand and the dramatic improve-
ments often induced by the presence of BF3. A good appli-
cation field of this concept is the enantioselective opening
of meso-epoxides. Indeed, many methodologies using asym-
metric catalytic systems have been successfully developed
for the enantioselective opening of meso-oxiranes by non-
carbon nucleophiles (including the CN group).[2] In con-
trast, there are only a few reports of efficient processes in-
volving organometallic species (RMgX, R2CuLi, R3Al,
RZnX etc.) for such a reaction, the best results being ob-
tained with organolithium reagents.[3–9] A significant ad-
vance was achieved in 1996 by Tomioka et al., who used an
organolithium/homochiral ether combination in the pres-
ence of BF3, as depicted in Scheme 1.[4,5]
Results and Discussion
Among the most usual chiral ligands used for organo-
lithium reagents, (–)-sparteine appears to be an excellent
typical diamine for such a study.[18] It is cheap, commer-
cially available, and very efficient in a number of reactions.
At first glance, it could seem incongruous to mix a strong
Lewis base (sparteine is a diamine) with a strong Lewis acid
such as BF3·Et2O as competition for the Lewis acid should
favor the diamine moiety instead of the oxygen of the epox-
ide. Despite this fact, we assumed that the kinetics of these
two competitive reactions (nucleophilic addition of the
organolithium reagents to the oxirane vs. irreversible com-
plexation of the Lewis acid and the diamine) might be sim-
ilar enough to afford at least small amounts of the desired
alcohols. We were thus very pleased to notice that when
1.5 equiv. of BF3·Et2O was added dropwise to an ethereal
solution of salt-free PhLi, complexed by 2 equiv. of (–)-
sparteine and 1 equiv. of cyclohexene oxide 1 (this stoichi-
ometry corresponds to that used by Tomioka et al.), a very
clean reaction took place to afford enantioenriched trans-
phenylcyclohexanol (2) in near quantitative yield (48%
ee;[19,20] Scheme 2).
Scheme 1.
[a] Laboratoire de Chimie Organique, UMR 7611, Université
Pierre et Marie Curie,
4 place Jussieu, 75252 Paris Cedex 05, France
E-mail: vrancken@ccr.jussieu.fr
[b] Département de chimie organique, Université de Genève
30 quai E. Ansermet, 1211 Genève 4, Switzerland
1354
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200400745
Eur. J. Org. Chem. 2005, 1354–1366

Enantioselective Desymmetrization of meso-Epoxides
FULL PAPER
and with a much lower selectivity (Table 2, Entry 6). This
result could be due to participation of the enantiomerically
enriched product in the enantioselective recognition. The
addition of varying catalytic amounts of enantiomerically
enriched trans-phenylcyclohexanol (2) to the reaction mix-
ture had no effect on the enantioselectivity, which rules out
the envisaged self-induction. The selectivity appears to de-
pend only on the BF3·Et2O/(–)-sparteine ratio, an optimum
result being reached for values above or equal to one
(Table 2, Entries 5–7).
Scheme 2.
Encouraged by this result, a study of the influence of
the different reaction parameters was initiated. It has been
reported in the literature that the temperature of formation
of the RLi/chiral ligand complex may have a dramatic effect
on its aggregation state, and thus affect the enantio-
selectivity.[21–23] However, we did not observe such an effect
− similar yields and enantioselectivities were obtained for
temperatures of formation of the (–)-sparteine/PhLi com-
plex ranging from –78 °C to 25 °C. On the other hand, the
addition of the Lewis acid at higher temperatures (–10 °C)
results in a dramatic drop in the selectivity (Table 1, En-
try 2). This is a general observation of our studies: the selec-
tivities were always strictly dependent on the temperature
of the reaction mixture. An intriguing solvent effect was
observed when hexane was used as cosolvent, both the
yields and selectivities decreasing proportionally to the hex-
ane/diethyl ether ratio (Table 1, Entries 3–5 vs. Entry 1).
Interestingly, the optimum selectivities were restored when
toluene was used instead of diethyl ether (Table 1, Entry 6).
As reported by Tomioka,[5] the selectivity was totally lost in
THF as its strong Lewis basicity might compete with the
diamine, thereby disrupting the formation of the sparteine/
PhLi complex (Table 1, Entry 7).
Table 2. Reaction of PhLi reagents/(–)-sparteine with cyclohexene
oxide and BF3·Et2O in diethyl ether or toluene/hexane mixtures as
solvent.
Entry
ArLi
RLi/sparteine/epoxide/ Yield [%] ee [%][b]
[a]
BF3
1
2
3
4
5
6
7
PhLi[c]
PhLi
PhLi
PhLi·LiBr
PhLi[c]
PhLi
2:2:1:1.5
2:4:1:1.5
2:1:1:1.5
2:4:1:1.5
2:2:1:5
2:2:1:0.5
1:1:1:0.5
Ͼ95
87
49
30
37
47
47
25
39
94
50[d]
Ͼ95
48
PhLi
49
[a] Isolated by column chromatography on silica gel. [b] Deter-
mined by 31P NMR spectroscopy according to ref.[19] [c] Salt-free
reagent. [d] Racemic trans-bromocyclohexanol was also obtained
in 17% yield.
Next, we screened a variety of organolithium reagents to
check quickly the scope of this reaction. As shown in
Table 3, only aryllithium reagents give significant ee values,
the use of alkyl-, alkenyl-, or alkynyllithium species giving
the corresponding alcohols with no selectivity (Entries 1–
6). However, the conversion is excellent for most of the
organolithium reagents tested.
Table 1. Solvent effect on enantioselectivity.
Entry
Solvent
T [°C]
Yield [%][a]
ee [%][b]
1
2
3
4
5
6
7
Et2O
Et2O
–78
–10
–78
–78
–78
–78
–78
Ͼ95
70
85
80
12
Ͼ95
Ͼ95
49
20
30
38
19
49
0
Table 3. Enantioselective nucleophilic ring-opening of 1.
Entry
RLi
Product
Yield [%][a]
ee [%][b]
Et2O/hexane[c]
Et2O/hexane[d]
hexane
1
2
3
4
5
6
Ph[c]
2
3
4
5
6
7
Ͼ95
97
49
0
6
5
0
Bu[c]
toluene/hexane[c]
THF
Me[c]
96
[d]
PhCH2
87[e]
98
1-cyclohexenyl[f]
1-hexynyl[f]
[a] Isolated by column chromatography on silica gel. [b] Deter-
mined by 31P NMR spectroscopy according to ref.[19] [c] Ratio: 5:1.
[d] Ratio: 10:1.
42[e]
0
[a] Isolated by column chromatography on silica gel. [b] Deter-
mined by 31P NMR spectroscopy according to ref.[19] [c] Salt-free
reagent. [d] Obtained by metalation with nBuLi. [e] This yield cor-
rects the previous reported one in ref.[6] [f] Obtained by halogen/Li
exchange with nBuLi.
Any variation of the number of equivalents of (–)-sparte-
ine with respect to the organolithium reagents gave lower
ee values. The optimum is a 1:1 ratio (Table 2, Entries 1–3).
If a PhLi·LiBr complex is used, additional (–)-sparteine is
needed, without enhancement of the ee. However, a lower
These results outline the surprising behavior of the Lewis
yield is observed due to the formation of bromocyclohexa- acid towards the diamine. This observation prompted us to
nol by competitive nucleophilic attack of the Br anion further study the addition of nBuLi to cyclohexene oxide 1
(Table 2, Entry 4). In the absence of BF3·Et2O, no reaction in the presence of BF3·Et2O and diamines as a model reac-
was observed at –78 °C, even if (–)-sparteine (Table 2, En- tion. First, the less bulky TMEDA was employed instead
try 5) is present. In this case, an increase of the temperature of (–)-sparteine, under standard reaction conditions. No ef-
to 0 °C for 3 d was needed to give a 34% yield of trans- fect on the yield was observed and the corresponding trans-
phenylcyclohexanol (2) with 5% ee. Similarly, a 1:2 BF3/ butylcyclohexanol (3) was obtained almost quantitatively
epoxide ratio allows the formation of 2 in only 49% yield, (Scheme 3).
Eur. J. Org. Chem. 2005, 1354–1366
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1355

E. Vrancken, A. Alexakis, P. Mangeney
FULL PAPER
Scheme 3.
Scheme 4.
Two possible explanations were proposed to explain this:
either there is no complexation of the BF3·Et2O with the
diamines (sparteine or TMEDA) in the reaction conditions,
or an equilibrium may allow the reaction to proceed. In
order to check these hypotheses, a precomplexation of
BF3·Et2O with the diamines was performed under different
conditions. The addition of BF3·Et2O to an ethereal solu-
tion of (–)-sparteine at room temperature leads to the exo-
thermic formation of a complex, as a white precipitate,
which is unable to achieve the transformation (Table 4, En-
try 1), which invalidates the hypothesis of an equilibrium.
It is noteworthy that the commercially available BF3·NEt3
reagent was also unable to promote the reaction, thus cor-
roborating this hypothesis (Table 4, Entry 5).The formation
of the Lewis acid/sparteine complex needed 1 h at low tem-
perature to ensure complete inhibition of the nucleophilic
ring-opening reaction (Table 4, Entries 2 and 3). The com-
plexation of BF3·Et2O to TMEDA is very fast even at
–78 °C (Table 4, Entry 4). The kinetics of the complexation
are thus strongly dependent on both the temperature and
the steric hindrance of the diamine.
Table 5. Ligand effects on enantioselectivity.
Entry Ligand RLi[a] Solvent
Pro-
duct
Yield [%][b] ee [%][c]
1[e]
2[e]
3[e]
4[e]
5[e]
6[e]
8a
8a
8b
8b
9
nBu tol/hex[d]
Ph tol/hex[d]
nBu tol/hex[d]
Ph tol/hex[d]
3
2
3
2
3
2
78
60
70
75
85
93
38
0
22
35
0
nBu
Et2O
Et2O
9
Ph
54
[a] Salt-free reagent. [b] Isolated by column chromatography on
silica gel. [c] Determined by 31P NMR spectroscopy according to
ref.[19] [d] In a 5:1 ratio. [e] RLi ligand/epoxide/BF3 ratio of
1.2:1.2:1:1.5.
ephedrine. Despite their similar structures, the use of these
two chiral ligands, in association with PhLi, resulted in dif-
ferent selectivities (Table 5, Entries 2 and 4). Moreover, the
trans-butylcyclohexanols 2 obtained with (1R,2S)-8a or
(1S,2S)-8b are of opposite absolute configuration, which
emphasizes the dramatic effect of the absolute configura-
tion of the benzylic carbon atom of these ligands (Table 5,
Entries 1 and 3). Although modest, the 38% ee obtained
with the nBuLi/(1R,2S)-8b complex is, to the best of our
knowledge, the best described in the literature. In the pres-
ence of the bis(oxazoline) 9, similar results to those ob-
served with (–)-sparteine were obtained: no selectivity with
nBuLi, and 54% ee with PhLi (Table 5, Entries 5 and 6).
Importantly, even this very strong Lewis base ligand is com-
patible with BF3 under our reaction conditions, as shown
by the high yield obtained.
Table 4. Effect of complexation of BF3·Et2O with diamines on nu-
cleophilic ring-opening of 1 by nBuLi.
Entry
Diamine
Temp [°C]
Time [min]
Yield [%][a]
1
2
3
sparteine
sparteine
sparteine
TMEDA
NEt3
25
–78
–78
–78
–
5
5
60
5
0
80
0
0
0
4
5[b]
–
[a] Determined by GC using decane as internal standard. [b] Com-
mercially available BF3·NEt3 was used.
In contrast, the complex formed when BF3·Et2O was
Since optically pure phenylcyclohexanol is a useful chiral
added to a preformed sparteine/nBuLi aggregate at –78 °C auxiliary,[24] we screened other aryllithium reagents which
was reactive, and subsequent addition of 1 after 5 or 30 min would afford other arylcyclohexanols. These compounds
yielding alcohol 3 cleanly. (Scheme 4). These observations may, in turn, serve as new chiral auxiliaries. The results are
suggest a possible protection of the diamine from the Lewis summarized in Table 6.
acid by the organolithium reagent.
All the arylcyclohexanol derivatives 10–19 obtained have
Despite the strong Lewis basicity of these two diamines, a (1R,2S) absolute configuration,[25] which results from a
the excellent yields obtained validate our approach, which nucleophilic attack on the (S)-carbon atom of the epoxide
allows the use of an extensive number of chiral auxiliaries. ring. A comparison of the results obtained with the three
We therefore decided to test other chiral ligands, such as isomers of tolyllithium shows that an ortho substituent
amino ethers 8a and 8b and bis(oxazoline) 9, as depicted in causes a marked increase in the enantioselectivity (Table 6,
Table 5.
Entry 1). However, an ortho-methoxy substituent has a det-
The two amino ethers 8a and 8b are easily obtained in rimental effect, probably due to internal coordination of the
three steps from (1R,2S)-ephedrine or (1S,2S)-pseudo- lithium cation by the oxygen atom (Table 6, Entry 4). The
1356
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1354–1366

Enantioselective Desymmetrization of meso-Epoxides
FULL PAPER
Table 6. Formation of arylcyclohexanols by reaction of aryllithiums
reagents with cyclohexene oxide.
We next examined the behavior of other epoxides with
PhLi and 1-naphthyllithium reagents. The absolute configu-
ration of the obtained alcohols could be determined by
derivatization with (+)- and (–)-methoxymandelic acid.[26]
The observed results indicate that, again, in every case the
nucleophilic attack takes place at the (S)-carbon atom of
the oxiranyl ring. As previously observed, 1-naphthyllith-
ium give better ee values than PhLi (Table 7, Entries 1, 2,
5, and 6). Although the ee values are lower than those ob-
tained with cyclohexene oxide (1), it is interesting to note
that a rather flat epoxide, such as cyclopentene oxide (20),
gave a result as good as cycloheptene oxide (23) or norbor-
nene oxide (25; Table 7, Entries 2, 6, and 8). Moreover, the
enantioselectivity depends on the substitution on the carbo-
cyclic ring: whereas the presence on the cyclopentene oxide
derivative 21 of a substituent anti to the epoxide ring de-
creases the ee from 53 to 24%, a higher ee (62%) is ob-
served for its syn isomer 22 (Table 7, Entries 2–4).
Table 7. Nucleophilic opening of various epoxides by an ArLi/spar-
teine/BF3·Et2O combination in toluene/hexane.
[a] Isolated by column chromatography on silica gel. [b] Deter-
mined by 31P NMR spectroscopy according to ref.[19] [c] Salt-free
reagent. [d] Obtained by halogen/Li exchange with nBuLi. [e] Ob-
tained by metalation with nBuLi. [f] In a 5:1 ratio. [g] ArLi/sparte-
ine/epoxide/BF3 ratio of 1:1:1:1. [i] ArLi/sparteine mixture insolu-
ble in the reaction conditions.
electronic nature of the para substituent plays a moderate
role, since the ee values are in the same range (43–53%).
The same may be roughly said about a meta substituent,
although we note the 60% ee obtained with a CF3 moiety
(Table 6, Entry 6). Both 1- and 2-naphthyllithium are ex-
tremely interesting cases (Table 6, Entries 8 and 9) as they
give the corresponding alcohols 17 and 18 with 70 and 85%
ee, respectively. Once again, as for the PhLi, an important
solvent effect is observed in diethyl ether/hexane mixtures
which disappears in toluene (see Table 6, Entries 4 and 9,
and Table 1). An ee of 87% was obtained in a large-scale
experiment (5 g of final material 18), showing that the
method has some preparative value. Moreover, enantiomer-
ically pure 18 (Ͼ98% ee) can be obtained in 78.5% yield
after one recrystallization. Similarly, compound 19 with
96% ee was obtained after one recrystallization. An in-
crease of the steric effects with anthracenyllithium (Table 6,
Entry 10) resulted in a lower ee.
[a] Isolated by column chromatography on silica gel. [b] Deter-
mined by 31P NMR spectroscopy according to ref.[19] [c] Salt-free
reagent. [d] Obtained by halogen/Li exchange with nBuLi. [e] Abso-
lute configuration (1R,2S) determined by polarimetry and com-
parison with the literature (see Exp. Sect.). [f] Absolute configura-
tion (1R,2S) estimated by comparison of the 31P NMR spectra of
the derivatives. [g] This value corrects the previous one reported in
ref.[6] [h] Absolute configuration (1R,2S) established according to
ref.[26] [i] Absolute configuration (1R,2S) established according to
ref.[26]
Eight-membered-ring epoxides are of particular interest.
We have already observed that the cyclooctene oxide 24
Eur. J. Org. Chem. 2005, 1354–1366
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1357

E. Vrancken, A. Alexakis, P. Mangeney
FULL PAPER
Scheme 5.
provides a mixture of the expected arylcyclooctanol 32 and to 84%. For our part, we decided to perform a brief study
the (–)-endo-cis-fused bicyclic alcohol 34, as a single diaster- of the influence of the different experimental parameters for
eoisomer, under our reaction conditions with PhLi.[27,28] the carbenoid rearrangement of epoxide 24 with the RLi/
This latter compound was first obtained by Cope,[29] start- (–)-sparteine/BF3 combination. The results are summarized
ing from the same epoxide and using strongly basic condi- in Table 8.
tions (LiNEt2, boiling Et2O, 48 h), and was rationalized as
The results obtained contrast with our observations
a result of an α-deprotonation followed by a transannular made for the nucleophilic ring-opening reaction of cyclo-
C–H insertion of the transient carbenoid species.[30–33] In hexene oxide (1). Indeed, similar ee values were obtained
our conditions at –78 °C, using stronger bases than PhLi for alcohol 34 either in Et2O/hexane or cumene/hexane mix-
and in diethyl ether instead of toluene, the reaction was tures (Table 8, Entries 1 and 2). Variation of the (–)-sparte-
chemoselective, yielding exclusively the rearranged alcohol ine/nBuLi ratio from 1:1 to 2:1 or of the BF3/(–)-sparteine
34 in quantitative yields (Scheme 5). As for the nucleophilic ratio from 2.5:1 to 0.25:1 have no significant effect on the
opening, the presence of the BF3 allows a strong accelera- enantioselectivity (Table 8, Entries 3–5). The use of LiBr as
tion of the carbenoid rearrangement, complete conversion additive has no beneficial influence on the ee values, ad-
being observed after the addition of the last drop of 1 equiv. ditional (–)-sparteine being needed in order to overcome the
of the Lewis acid. Presumably, this accelerating effect can 1:1 Li cation/ligand ratio (Table 8, Entries 6 and 7).
be interpreted in this case by the higher acidity of the oxir-
The reverse addition of epoxide 24 to a preformed mix-
anyl protons, or by an enhanced carbenoid character of the ture of nBuLi/sparteine complex and BF3·Et2O gave
lithiooxirane intermediate, or both. Note that the RLi/(–)- alcohol 34 with no change in the selectivity (Table 8, En-
sparteine complex causes the α-deprotonation of the (R)- try 1), which points to the crucial role of the Lewis acid in
carbon atom of the oxiranyl ring, whereas the (S) center is this enantioselective process (Scheme 6).
preferentially attacked during nucleophilic opening. This
fact may be due to a difference between the transition states increasing the steric hindrance of the organolithium used
of these two processes. (48% for nBuLi to 61% for sBuLi; Table 8, Entries 1, 9, and
A slight enhancement of the ee values was observed by
An enantioselective version of this reaction has been re- 10). These values are slightly different to Hodgson’s results:
ported by Hodgson et al., using alkyllithium bases in the they are higher for nBuLi (48% vs. 31%) and lower for the
presence of (–)-sparteine.[34–36] Under their optimum condi- secondary organolithium reagents (61% vs. 77% for sBuLi
tions (iPrLi, Et2O, –98 °C, 6 h to room temperature, 12 h) and 56% vs. 84% for iPrLi).[34–36] In our case, sBuLi gave
the alcohol 34 was obtained in good yield with an ee of up the best selectivity, whereas without the assistance of BF3,
Table 8. Influence of some reaction parameters on yield and enantioselectivity.
Entry
RLi
RLi/sparteine/epoxide/BF3
Solvent
Yield [%][a]
ee [%][b]
1
2
3
4
5
6
7
8
9
nBuLi
nBuLi
nBuLi
nBuLi
nBuLi
2:2:1:1.5
2:2:1:1.5
2:4:1:1.5
2:2:1:0.5
2:2:1:5
2:2:1:1.5
2:4:1:1.5
2:2:1:1.5
2:2:1:1.5
2:2:1:1.5
Et2O/hexane[c]
cumene/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
Et2O/hexane[c]
98
60
85
48
45
48
48
47
27
45
–
45
98
nBuLi, LiBr
nBuLi, LiBr
MeLi
78[e]
72[e]
0
iPrLi
sBuLi
96
98
56
61
10
[b]
[a] Isolated by column chromatography on silica gel. Determined by 31P NMR spectroscopy according to ref.[19] [c] In a 6:1 ratio. [d]
Racemic trans-bromocyclooctanol was also obtained in 20% yield.
1358
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1354–1366

Enantioselective Desymmetrization of meso-Epoxides
FULL PAPER
Scheme 6.
as reported by Hodgson, iPrLi gave the best ee. Therefore,
Under the same conditions, but in the presence of
the BF3 not only has an influence on the kinetics of the BF3·Et2O, the course of the reaction changes dramatically:
reaction, but also on the enantioselectivity. This effect illus- the major compound, obtained in 90% yield, is the alcohol
trates possible differences in the transition states of theses 37, as a single diastereoisomer and with an interesting 64%
two processes. Indeed, the α-deprotonation is known to ee. As previously, the absolute configuration could be as-
proceed through the complexation of the organolithium sessed by derivatization with (+)- and (–)-methoxymandelic
with the oxirane, the deprotonation taking place into an acids,[26] and was found to be (1S,2R,3S), which is in good
epoxide/RLi oligomer.[30,31,33,37] The formation of this agreement with an abstraction of the α-proton from the (R)-
oligomeric structure should be strongly disrupted by the carbon atom of the oxiranyl ring. Moreover, the obtention
complexation of the Lewis acid to the oxygen atom of the of a single diastereoisomer supports the hypothesis of a
oxiranyl ring, thus inducing a probable intermolecular tran- concerted carbenoid rearrangement. These observations are
sition state.
in good accordance with an enhancement of the oxiranyl
The presence of the BF3 and the basicity of the organo- proton acidity induced by the complexation with the Lewis
lithium reagent used can thus have a profound effect on the acid.
course of the reaction. In all cases, a strong Lewis base can
Besides the major product 37, trace amounts (9%) of sec-
be used as chiral ligand without any erosion of the conver- butyl-substituted cyclooctanol 38 were detected. Such a
sion, and, in this way, various valuable enantiomerically en- compound, which results from a nucleophilic ring-opening
riched products can be synthesized. The application of this reaction, has never been observed when cyclooctene oxide
methodology to the synthesis of meso-cycloocta-1,5-diene 24 is submitted to the same reaction conditions. The forma-
oxide (35) illustrates the above concept well. As shown in tion of this side product is consistent with the above hy-
Scheme 7, addition of 35 to a 1:1 sBuLi/(–)-sparteine mix- pothesis upon the effect of the double bond, which probably
ture at –90 °C gives, after 6 h, the allylic alcohol 36 in 85% affects the oxiranyl proton acidity. To check it, the less basic
yield, with an ee of up to 62% (the absolute configuration PhLi was used instead of sBuLi under otherwise similar
was not determined). Note that under the same reaction conditions, and, as expected, the nucleophilic ring-opening
conditions, its saturated analog 24 gave exclusively the α- product 39 was obtained in 81% yield with an ee of up to
deprotonation product 34. This difference of behavior has 51%. The absolute configuration, assigned after hydrogena-
been rationalized by Crandall[38] as a result of a steric de- tion of the double bond and comparison of the optical rota-
congestion of the carbocyclic ring, induced by the presence tion of the resulting product with alcohol 32, was found to
of the double bond, which may increase the flexibility and be (1R,2S).
allows a suitable conformation for the postulated β-elimi-
In order to improve the ee of bicyclic alcohol 37, a slow
nation process.[39] In the same manner, this steric deconges- (30 min) addition of BF3 with a syringe pump was realized
tion could reduce the strain energy of the epoxide ring, thus in a scaled-up trial (20 mmol). A slight enhancement of the
decreasing the acidity of the α-proton.[32]
selectivity, from 64 to 75% ee, was observed, which could
Scheme 7.
Eur. J. Org. Chem. 2005, 1354–1366
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1359

E. Vrancken, A. Alexakis, P. Mangeney
FULL PAPER
Scheme 8.
be explained by a better control of the reaction temperature. quently, when the epoxide is treated simultaneously with
However, the allylic alcohol 36 appears as a new side pro- RLi and BF3, the α-deprotonation step and the transannu-
duct. To avoid its formation, the Lewis acid was added lar rearrangement may be compressed into a single, con-
prior to the oxirane, giving the alcohol 37 cleanly with a certed process, thus avoiding the formation of a real car-
non-optimized 71% ee (Scheme 8).
benoid intermediate [Scheme 10, Equation (2)]. This type of
Finally, we tried to extend our methodology to the highly model could be applicable to all medium-sized-ring epox-
reactive endo-norbornene oxide 25. This epoxide is known ides to explain their observed behavior towards our RLi/
to give the 1,3-carbenoid C–H insertion product (alcohol sparteine/BF3 combination.
40) upon treatment with organolithium reagents
(Scheme 9).[36,40–42]
Conclusions
We have demonstrated that strong Lewis bases, such as
(–)-sparteine, are compatible with strong Lewis acids, such
as BF3, as long as they are protected by coordination to an
organolithium reagent. Moreover, when the Lewis base is
Scheme 9.
chiral, good enantioselectivity can be attained in the nucleo-
philic desymmetrization of meso-epoxides. The role of the
Lewis acids has been demonstrated to be exclusively di-
rected towards the activation of the epoxides, to enhance
the reaction rate.
An attempt to develop an enantioselective version of this
reaction with our RLi/sparteine/BF3 combination, under
standard conditions, resulted in the degradation of the
starting material. However, a 70% yield of the desired
alcohol 40 was obtained when the Lewis acid was added
to the sBuLi/(–)-sparteine mixture before the oxirane, but
unfortunately without selectivity. These results highlight the
mechanism of the base-promoted rearrangement of me-
dium-sized-ring epoxides in the presence of BF3. The en-
hanced acidity of endo-norbornene oxide 25 is well
known.[32] Therefore, under the standard conditions, the
formation of the transient carbenoid species should occur
before the addition of BF3. The obtention of degradation
products would thus reflect the instability of the carbenoid
Experimental Section
General Considerations: Experiments involving organometallic rea-
gents were carried out under dry argon. All glassware was dried at
120 °C and assembled while hot under a stream of dry nitrogen.
All moisture-sensitive reactants were handled under argon. Low-
temperature experiments were carried out by cooling a three-
necked, round-bottomed flask with an acetone (–80 °C) or diethyl
ether (–95 °C) bath frozen with liquid nitrogen. The flask was
equipped with an internal thermometer, an argon inlet, and a sep-
in the presence of BF3 [Scheme 10, Equation (1)]. Conse- tum cap. Diethyl ether and tetrahydrofuran were distilled from so-
Scheme 10.
1360
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1354–1366

Enantioselective Desymmetrization of meso-Epoxides
FULL PAPER
dium/benzophenone ketyl. Organolithium reagents were titrated
with 2-butanol (1 m solution in toluene) using 1,10-phenanthroline
as indicator. (–)-Sparteine was freshly distilled from calcium hy-
dride before use. Column chromatography was performed on silica
gel Si 60 (0.015–0.040 or 0.040–0.063 mesh). Thin-layer chromatog-
of potassium hydride (2.2 g, 55 mmol) in THF (200 mL). The mix-
ture was stirred for 16 h, after which a solution of MeI (5.08 g,
35.5 mmol) in THF (50 mL) was added. After a further 3 h, the
reaction was carefully quenched by addition of ice-cold water
(400 mL), and the aqueous layer extracted with diethyl ether
raphy (TLC) was performed on silica gel (0.25 mm, F-254) and (3×150 mL). The organic layer was dried with K2CO3 and filtered.
visualized under a UV lamp (254 and 366 nm) or by spraying with
a 10% phosphomolybdic acid solution in ethanol (heating), vanillin
reagent (heating), or Dragendorf reagent. Optical rotations were
measured at 20 °C. 1H NMR spectra were recorded at 200 or
400 MHz, 13C NMR spectra were recorded at 50 or 100 MHz, 31P
NMR spectra were recorded at 162 MHz, and 19F NMR spectra
were recorded at 376 MHz, in CDCl3 as solvent. Chemical shifts
The solvent was removed under reduced pressure and the residue
was chromatographed on silica gel using cyclohexane/ethyl acetate
(50:50) as eluent to afford the methoxy-acetamide derivative
1
(5.02 g, 55%) as a white solid. H NMR (400 MHz, CDCl3): δ =
7.33 (m, 5 H), 4.21 (m, 1 H), 4.07 (m, 1 H), 3.99 (m, 1 H), 3.39,
3.38, 3.15, and 3.12 (4s, 6 H), 2.87 and 2.91 (2s, 3 H, H6), 0.98 (d,
J = 5.8 Hz) ppm. 13C NMR (100 MHz, CDCl3): δ = 170.6, 169.6,
139.5, 139.3, 129.0, 128.8, 128.7, 127.9, 127.8, 85.1, 84.9, 72.2, 59.3,
59.1, 57.2, 56.7, 27.3, 15.9, 14.3 ppm. The methoxy acetamide de-
rivative (5.02 g, 20 mmol) was slowly added to a solution of LiAlH4
(1.52g, 40 mmol) in diethyl ether (120 mL) and the reaction mixture
was heated under reflux for 3 d. The hydrolysis was carried out at
0 °C with an aqueous solution of sodium/potassium tartrate
(100 mL). The mixture was stirred for 16 h and then the aqueous
layer was extracted with diethyl ether (3×50 mL). The combined
organic layers were dried with K2CO3, filtered, and the solvent was
removed under reduced pressure. The product was purified by col-
umn chromatography on silica gel with cyclohexane/ethyl acetate/
NEt3 (50:45:5) as eluent to afford amino ether 8b (3.79 g, 80%) as
a colorless oil. 1H NMR (200 MHz, CDCl3): δ = 7.31 (m, 5 H),
4.02 (d, J = 8.9 Hz, 1 H), 3.50 (t, J = 6.3 Hz, 2 H), 3.36 (s, 3 H),
3.13 (s, 3 H), 2.97 (m, 1 H), 2.86 (m, 1 H), 2.75 (m, 1 H), 2.42 (s,
3 H), 0.67 (d, J = 6.8 Hz, 3 H) ppm. 13C NMR (100 MHz, CDCl3):
δ = 140.9, 128.3, 127.8, 126.7, 86.5, 71.8, 63.6, 59.9, 56.4, 52.9,
39.0, 11.9 ppm. C14H23NO2: calcd. C 70.85, H 9.77, N 5.90; found
C 70.89, H 9.81, N 5.72. [α]2D0 = +86.05 (c = 1.45, CHCl3).
1
are reported in ppm (reference CDCl3 for H and 13C, H3PO4 for
31P and CFCl3 for 19F).
Preparation of the Epoxides: All the epoxides have already been
described. Epoxides 1, 20, and 23–25 are commercially available.
Epoxide 35 was prepared according to the procedure of Crandall
et al.[43] Epoxides 21 and 22 were prepared according to the pro-
cedure of Asami.[44]
(2-Methoxyethyl)[(1R,2S)-2-methoxy-1-methyl-2-phenylethyl]meth-
ylamine (1R,2S)-8a: A solution of methoxyacetyl chloride
(1.53 mL, 16.85 mmol) in dichloromethane (10 mL) was added at
0 °C under nitrogen to a solution of (1R,2S)-O-methylephedrine[45]
(3 g, 16.9 mmol) and triethylamine (7 mL, 50 mmol) in dichloro-
methane (60 mL). The reaction mixture was stirred at room tem-
perature for 16 h and then poured into a saturated solution of
NH4Cl. The aqueous layer was extracted with dichloromethane
(2×50 mL) and the combined organic phases were washed with
brine, dried with K2CO3, and filtered. The solvents were removed
under reduced pressure to afford the crude product (4.03 g, 95%)
which was then slowly added without further purification to a solu-
tion of LiAlH4 (1.25 g, 33.7 mmol) in diethyl ether (100 mL). The
reaction mixture was heated under reflux for 3 d and was then hy-
drolyzed at 0 °C with an aqueous solution of sodium/potassium
tartrate (100 mL). The mixture was stirred for 16 h and then the
aqueous layer was extracted with diethyl ether (3 ×50 mL). The
combined organic layers were dried with K2CO3, filtered, and the
solvent was removed under reduced pressure. The product was
purified by column chromatography on silica gel with cyclohexane/
ethyl acetate/NEt3 (50:45:5) as eluent to afford the title compound
(2.94 g, 11.66 mmol, 69 % 2 steps) as a colorless oil. 1H NMR
(200 MHz, CDCl3): δ = 7.30 (m, 5 H), 4.32 (d, J = 3.95 Hz, 1 H),
3.40 (t, J = 11 Hz, 2 H), 3.23 and 3.34 (2 s, 6 H), 2.76 (m, 3 H),
2.36 (s, 3 H), 1.02 (d, J = 6.8 Hz, 3 H) ppm. 13C NMR (100 MHz,
CDCl3): δ = 141.6, 128.1, 127.0, 126.8, 85.0, 71.6, 64.6, 58.8, 56.7,
53.0, 39.4, 8.8 ppm. C14H23NO2: calcd. C 70.85, H 9.77, N 5.90;
found C 70.77, H 9.75, N 5.90. [α]2D0 = –53.08 (c = 1.32, CHCl3).
General Procedure A. Nucleophilic Ring-Opening of meso-Epoxides
by Organolithium/(–)-Sparteine/BF3·Et2O: A solution of salt-free
RLi (4 mmol) was added dropwise, under argon, to a cooled
(–78 °C) solution of (–)-sparteine (0.92 mL, 4 mmol) in the required
solvent (12 mL). After stirring for 30 m at this temperature, the
epoxide (2 mmol) was injected with a syringe. A solution of
BF3·Et2O (0.38 mL, 3 mmol) in the appropriate solvent (2 mL) was
slowly (over 5 min) added in order to maintain the temperature at
–78 °C. After stirring for 10 min, the reaction was quenched with
MeOH (2 mL) and NEt3 (3 mL) and then the mixture was allowed
to reach room temperature. An aqueous solution of 5% H2SO4
(10 mL) was then slowly added to the reaction mixture. The aque-
ous layer was extracted with Et2O (3×50 mL), the combined or-
ganic phases were washed with brine, dried with MgSO4, and fil-
tered. The solvents were removed under reduced pressure and the
product was then subjected to flash chromatography (FC) on SiO2.
The enantiomeric excess of the compounds was determined by 31P
NMR spectroscopy, as reported previously.[19]
(2-Methoxyethyl)[(1S,2S)-2-methoxy-1-methyl-2-phenylethyl]meth-
ylamine (1S,2S)-8b: A solution of methoxyacetyl chloride (5.56 mL,
60 mmol) in dichloromethane (30 mL) was added dropwise over
30 min, in order to keep the temperature at –20 °C, to a solution
of (1S,2S)-(+)-pseudoephedrine (9.9 g, 60 mmol) and triethylamine
(11.15 mL, 80 mmol) in dichloromethane (130 mL). The reaction
mixture was stirred at this temperature for 2 h and then poured
into a saturated solution of NH4Cl. The aqueous layer was ex-
tracted with dichloromethane (2×50 mL) and the combined or-
ganic phases were washed with brine, dried with MgSO4, and fil-
tered off. The solvents were removed under reduced pressure to
afford the acetamide derivative as a white solid (9.95 g) which was
used without further purification. A solution of the acetamide
(8.40 g) in THF (70 mL) was added dropwise over 1 h to a slurry
(1R,2S)-trans-2-Phenyl-1-cyclohexanol (2):[5] The product was iso-
lated by FC (pentane/diethyl ether, 80:20) as a white solid. The
NMR spectra of 2 are identical to those reported previously.[5] 31
P
NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-
ethanediamine, CDCl3]: δ = 141.60 (s, 74.5%), 142.98 (s, 25.5%)
ppm. ee = 49%. [α]2D0 = –24 (c = 1.26, benzene) for 49% ee {ref.[5]
[α]2D0 = –22.4 (c = 1.26, benzene) for 47% ee}. M.p. 57 °C (ref.[46]
56–57 °C).
trans-2-Butyl-1-cyclohexanol (3):[5] The product was isolated by FC
(pentane/diethyl ether, 85:15) as a colorless oil. The NMR spectra
of 3 are identical to those reported previously.[5] 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
Eur. J. Org. Chem. 2005, 1354–1366
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1361

E. Vrancken, A. Alexakis, P. Mangeney
FULL PAPER
amine, CDCl3]: δ = 143.38 (s, 50%), 140.07 (s, 50%) ppm. ee = 0%. MeOH) for 52% ee {ref.[51] [α]2D0 = –59.5 (c = 1.37, MeOH) for
99% ee}. M.p. 70–71 °C (ref.[53] 72–73 °C).
trans-2-Methyl-1-cyclohexanol (4):[47] The product was isolated by
(1R,2S)-trans-2-(2-Methoxyphenyl)-1-cyclohexanol (13): The pro-
duct was isolated by FC (pentane/diethyl ether, 70:30) as a pale-
yellow solid. 1H NMR (400 MHz, CDCl3): δ = 7.25–6.97 (m, 4 H),
3.83 (s, 3 H), 3.75 (m, 1 H), 3.03 (m, 1 H), 2.16 (m, 1 H), 1.78 (m,
4 H), 1.44 (m, 4 H) ppm. 13C NMR (200 MHz, CDCl3): δ = 159.8,
131.5, 127.3, 120.9, 111.9, 73.8, 55.4, 45.9, 35.1, 32.3, 26.1,
25.1 ppm. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-di-
phenyl-1,2-ethanediamine, in the presence of S8, CDCl3]: δ = 82.25
(s, 57.5%), 82.13 (s, 42.5%) ppm. ee = 15%. [α]2D0 = –5 (c = 1.52,
FC (pentane/diethyl ether, 85:15) as a colorless oil. The NMR spec-
tra of 4 are identical to those reported previously.[47] 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, CDCl3]: δ = 143.32 (s, 51.5%), 140.14 (s, 48.5%) ppm. ee =
3%.
trans-2-Benzyl-1-cyclohexanol (5):[48] The product was isolated by
FC (pentane/diethyl ether, 80:20) as a white solid. 1H NMR
3
(400 MHz, CDCl3): δ = 7.28 (m, 5 H), 3.23 (m, 1 H), 3.10 (dd, J
2
3
2
= 3.9, J = 13 Hz, 1 H), 2.38 (dd, J = 9.2, J = 13 Hz, 1 H), 1.99
(m, 1 H), 1.54 (m, 6 H), 1.12 (m, 1 H), 0.93 (m, 1 H) ppm. 13C
NMR (100 MHz, CDCl3): δ = 140.7, 129.4, 128.1, 74.5, 125.7, 47.0,
39.0, 35.8, 29.9, 25.4, 24.9 ppm. 31P NMR [162 MHz, (1R,2R)-(+)-
N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 143.55
(s, 50%), 140.46 (s, 50%) ppm; ee = 0%. M.p. 76 °C (ref.[48] 76–
77 °C).
MeOH) for 15% ee. IR: ν = 3392, 2929, 2853, 1599, 1492, 1461,
˜
1240, 1052, 752 cm–1. M.p. 50–51 °C. C13H18 O2: calcd. C 75.69, H
8.80; found C 75.68, H 8.76.
(1R,2S)-trans-2-(4-Methoxyphenyl)-1-cyclohexanol (14):[51] The
product was isolated by FC (pentane/diethyl ether, 70:30) as a white
1
solid. H NMR (400 MHz, CDCl3): δ = 7.28 (m, 2 H), 6.91 (m, 2
H), 1.40 (m, 3 H), 3.83 (s, 3 H), 3.62 (m, 1 H), 2.39 (ddd, J = 3.5,
9.9, 12.2 Hz, 1 H), 2.15 (m, 1 H), 1.87 (m, 2 H), 1.85 (m, 1 H),
1.52 (m, 2 H) ppm. 13C NMR (200 MHz, CDCl3): δ = 158.8, 135.5,
129.1, 114.5, 74.9, 55.5, 52.7, 34.7, 33.8, 26.5, 25.4 ppm. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, CDCl3]: δ = 141.65 (s, 72%), 142.98 (s, 28%) ppm. ee =
trans-2-(1-Cyclohexenyl)-1-cyclohexanol (6):[49] The product was
isolated by FC (pentane/diethyl ether, 85:15) as a colorless oil. H
1
NMR (400 MHz, CDCl3): δ = 5.59 (br. s, 1 H), 3.40 (dt, J = 4.2,
10 Hz, 1 H), 1.23 (m, 5 H), 2.03 (m, 3 H), 1.92 (m, 3 H), 1.66 (m,
6 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 138.7, 124.9, 70.8,
55.4, 34.3, 30.2, 35.2, 26.2, 25.6, 23.3, 23.0 ppm. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, in the presence of S8, CDCl3]: δ = 82.60 (s, 50%), 81.94 (s,
50%) ppm. ee = 0%.
44%. [α]2D0 = –24 (c = 1.46, MeOH) for 44% ee {ref.[51] [α]2D0
=
–55.4 (c = 1.46, MeOH) for 99% ee}. M.p. 70–71 °C (ref.[47] 70–
71 °C).
(1R,2S)-trans-2-(3-Trifluoromethylphenyl)-1-cyclohexanol (15): The
trans-2-(1-Hexynyl)-1-cyclohexanol (7):[50] The product was isolated
by FC (pentane/diethyl ether, 90:10) as a colorless oil. The NMR
spectra of 7 are identical to those reported previously.[50] 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, CDCl3]: δ = 140.91 (s, 50%), 139.71 (s, 50%) ppm. ee = 0%.
product was isolated by FC (pentane/diethyl ether, 80:20) as a white
1
solid. H NMR (400 MHz, CDCl3): δ = 7.53 (m, 2 H), 7.46 (m, 2
H), 3.62 (m, 1 H, H1), 2.54 (ddd, J = 3.4, 10.1, 12.3 Hz, 1 H, H2),
2.16 (m, 2 H), 1.80 (m, 2 H), 1.60 (m, 2 H), 1.43 (m, 2 H) ppm.
13C NMR (100 MHz, CDCl3): δ = 144.7, 131.4, 129.1, 124.5, 123.7,
74.2, 53.0, 34.9, 33.4, 25.9, 25.0 ppm. 19F NMR (376 MHz,
CDCl3): δ = –62.97 ppm. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-
dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 143.78 (s,
80%), 141.81 (s, 20%) ppm. ee = 60%. [α]2D0 = –15 (c = 1.49,
(1R,2S)-trans-2-(2-Methylphenyl)-1-cyclohexanol (10):[51] The pro-
duct was isolated by FC (pentane/diethyl ether, 85:15) as a colorless
1
oil. H NMR (400 MHz, CDCl3): δ = 7.21 (m, 4 H), 4.02 (dt, J =
4.3, 9.7 Hz, 1 H), 2.81 (ddd, J = 3.4, 10, 12 Hz, 1 H), 2.42 (s, 3 H),
2.14 (m, 1 H), 1.86 (m, 2 H), 1.83 (m, 2 H), 1.41 (m, 4 H) ppm.
13C NMR (200 MHz, CDCl3): δ = 141.4, 137.1, 130.5, 126.4, 126.1,
125.4, 74.3, 47.7, 34.5, 33.1, 26.2, 25.1, 19.8 ppm. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, CDCl3]: δ = 143.22 (s, 82.5%), 141.67 (s, 17.5%) ppm. ee =
65%. [α]2D0 = –42 (c = 1.28, CHCl3) for 65% ee {ref.[51] [α]2D0 =–63.9
(c = 1.26, benzene) for 90% ee}.
(1R,2S)-trans-2-(3-Methylphenyl)-1-cyclohexanol (11):[52] The pro-
duct was isolated by FC (pentane/diethyl ether, 80:20) as a colorless
oil. 1H NMR (400 MHz, CDCl3): δ = 7.08 and 7.26 (2 m, 4 H),
3.67 (dt, J = 4.2, 10 Hz, 1 H), 2.40 (m, 1 H), 2.37 (s, 3 H), 2.15 (m,
1 H), 1.86 (m, 2 H), 1.77 (m, 1 H), 1.61 (m, 1 H), 1.55 (m, 1 H),
1.43 (m, 3 H) ppm. 13C NMR (200 MHz, CDCl3): δ = 143.0, 138.3,
128.6, 127.5, 124.8,. 74.3, 53.1, 34.3, 33.2, 26.0, 25.0, 21.4 ppm.
31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-
ethanediamine, CDCl3]: δ = 141.35 (s, 32%), 143.18 (s, 68%) ppm.
ee = 36%. [α]2D0 = –9 (c = 0.90, CHCl3) for 36% ee.
MeOH) for 60% ee. M.p. 53–54 °C. IR: ν = 3366, 2929, 1149, 1319,
˜
1102, 801, 701 cm–1. C13H15 F3O: calcd. C 63.93, H 6.19; found C
63.89, H 6.17.
(1R,2S)-trans-2-(4-Fluorophenyl)-1-cyclohexanol (16): The product
was isolated by FC (pentane/diethyl ether, 80:20) as a white solid.
1H NMR (400 MHz, CDCl3): δ = 7.25 (m, 2 H), 7.04 (m, 2 H),
3.62 (m, 1 H), 2.44 (ddd, J = 3.5, 9.9, 12.0 Hz, 1 H), 2.13–1.46 (m,
8 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 163.3, 161.0, 139.3,
129.6, 129.5, 116.0, 115.8, 74.9, 53.8, 34.9, 33.8, 26.3, 25.4 ppm.19F
NMR (376 MHz, CDCl3): δ = –116.8 ppm. 31P NMR [162 MHz,
(1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine,
CDCl3]: δ = 143.28 (s, 76.5%), 141.71 (s, 23.5%) ppm. ee = 53%.
[α]2D0 = –12.3 (c = 1.52, MeOH) for 53% ee. M.p. 63–64 °C. IR: ν
˜
= 3319, 2935, 1064, 1509, 1227, 1059, 810 cm–1. C12H15FO: calcd.
C 74.20, H 7.78; found C 74.18, H 7.74.
(1R,2S)-trans-2-(2-Naphthyl)-1-cyclohexanol (17):[5] The product
was isolated by FC (pentane/diethyl ether, 80:20) as a white solid.
1H NMR (200 MHz, CDCl3): δ = 7.30–7.90 (m, 7 H), 3.62 (m, 1
H), 2.53 (m, 1 H), 1.30–2.30 (m, 8 H) ppm. 13C NMR (100 MHz,
CDCl3): δ = 133.6, 132.6, 128.4, 127.6, 127.6, 126.7, 126.1,125.9,
125.5, 74.2, 53.3, 34.5, 33.3, 26.0, 25.1 ppm. 31P NMR [162 MHz,
(1R,2S)-trans-2-(4-Methylphenyl)-1-cyclohexanol (12):[51] The pro-
duct was isolated by FC (pentane/diethyl ether, 80:20) as a white
solid.1H NMR (400 MHz, CDCl3): δ = 7.15 (m, 4 H), 3.67 (m, 1
H), 2.41 (ddd, J = 3.5, 9.9, 12.3 Hz, 1 H), 2.35 (s, 3 H), 2.14 (m, 1
H), 1.86 (m, 2 H), 1.71 (m, 1 H), 1.39 (m, 5 H) ppm. 13C NMR (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine,
(200 MHz, CDCl3): δ = 140.4, 136.3, 129.4, 127.8, 74.3, 52.7, 34.4, CDCl3]: δ = 142.59 (s, 84.9%), 140.54 (s, 15.1%) ppm. ee = 70%.
33.4, 26.0, 25.0, 21.0 ppm. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ- [α]2D0 = –20.2 (c = 1.035, CHCl3) for 70% ee {ref.[5] [α]2D0 = –7.8 (c
dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 143.18 (s,
= 1.05, CHCl3) for 33% ee}. M.p. 89 °C (2-propanol) (ref.[54] 90 °C,
76%), 141.58 (s, 24%) ppm. ee = 52%. [α]2D0 = –31 (c = 1.28,
2-propanol). Recrystallization from EtOH afforded white crystals
1362
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1354–1366

Enantioselective Desymmetrization of meso-Epoxides
FULL PAPER
of racemic material. Subsequent filtration and concentration of the
mother liquor under reduced pressure gave an enantiomerically en-
riched solid (96% ee). The enantiomeric excess was determined by
31P NMR spectroscopy as well as by HPLC (Chiral OD; iPrOH/
H), 1.85 (m, 1 H), 0.95 (s, 9 H), 0.12 (s, 6 H) ppm. 13C NMR
(100 MHz, CDCl3): δ = 139.6, 134.3, 132.7, 129.2, 126.2, 126.1,
125.8, 124.1, 78.8, 71.5, 48.3, 44.7, 42.9, 26.3, –4.3. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
hexane, 2:98; 1 mLmin–1; 254 nm). 31P NMR [162 MHz, (1R,2R)- amine, CDCl3]: δ = 142.43 (s, 62%), 137.09 (s, 38%) ppm. ee =
(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ =
142.59 (s, 97.9%), 140.54 (s, 2.1%) ppm. ee = 96%. [α]2D0 = –25.6
(c = 1.01, CHCl3) for 96% ee.
24%. [α]2D0 = +0.8 (c = 1.22, CHCl ) for 24% ee. IR: ν = 3380,
˜
3
3050, 2959, 1596, 1471, 1047, 777 cm–1. C21H30O2Si: calcd. C 73.63,
H 8.20; found C 73.40, H 7.90. Derivatization of alcohol with (R)-
(–)-2-methoxy-2-phenylacetic acid: 1H NMR (400 MHz, CDCl3): δ
= 3.91 (dt, J = 5.8, 8.6 Hz, 41%), 3.81 (dt, J = 5.8, 8.6 Hz, 59%),
3.26 (s, 61%), 3.22 (s, 39%) ppm. Derivatization of alcohol with
(S)-(+)-2-methoxy-2-phenylacetic acid: 1H NMR (400 MHz,
CDCl3): δ = 3.98 (dt, J = 5.8, 8.6 Hz, 60.5%), 3.84 (dt, J = 5.8,
8.6 Hz, 39.5%), 3.31 (s, 38.5%), 3.27 (s, 61.5%) ppm.
(1R,2S)-trans-2-(1-Naphthyl)-1-cyclohexanol (18):[5] The product
was isolated by FC (pentane/diethyl ether, 80:20) as a white solid.
1H NMR (200 MHz, CDCl3): δ = 8.22–7.55 (m, 7 H), 4.02 (m, 1
H), 3.51 (br. s, 1 H), 2.27 (br. s, 1 H), 1.97 (m, 2 H), 1.85 (m, 1 H),
1.59 (m, 6 H) ppm. 13C NMR (200 MHz, CDCl3): δ = 139.6, 134.2,
132.7, 129.0, 126.1, 125.7, 125.8, 123.3, 74.3, 46.7, 34.8, 34.0, 26.5,
25.2 ppm. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-di-
phenyl-1,2-ethanediamine, CDCl3]: δ = 143.41 (s, 92.5%), 140.85
(s, 7.5%) ppm. ee = 85%. [α]2D0 = –60 (c = 1.46, MeOH) for 85%
ee {ref.[51] [α]2D0 = –72.9 (c = 1.47, MeOH,) for 99% ee}. M.p.129–
130 °C (ref.[55] 129–130 °C). Recrystallization from EtOH afforded
white crystals of racemic material. Subsequent filtration andcon-
centration of the mother liquor under reduced pressuregave an en-
antiomerically enriched solid (ee Ͼ 98%). 31P NMR [162 MHz,
(1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine,
CDCl3]: δ = 142.59 (s, Ͼ98%), 140.54 (s, Ͻ2%) ppm. ee Ͼ 98%.
[α]2D0 = –71.5 (c = 1.38, MeOH) for 98% ee.
(1R,2S,5R)-trans-(4-tert-Butyldimethylsiloxy)-2-(1-naphthyl)-1-cy-
clopentanol (29): The product was isolated by FC (pentane/diethyl
ether, 80:20) as acolorless oil. 1H NMR (400 MHz, CDCl3): δ =
8.37–7.25 (m, 7 H), 4.47 (m, 1 H), 4.17 (m, 1 H), 4.11 (dt, J = 3,
8.5 Hz, 1 H), 2.43–2.01 (m, 4 H), 0.98 (s, 9 H), 0.16 (s, 6 H) ppm.
13C NMR (100 MHz, CDCl3): δ = 140.2, 134.3, 132.7, 129.1, 126.4,
125.7, 124.5, 122.2, 80.4, 74.8, 49.1, 44.0, 42.5, 26.2, –4.3 ppm. 31P
NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-eth-
anediamine, CDCl3]: δ = 142.95 (s, 81%), 139.28 (s, 19%) ppm. ee
= 62%. [α]2D0 = –1.5 (c = 1.27, CHCl ) for 62% ee. IR: ν = 3600,
˜
3
3051, 2889, 1598, 1471, 1063, 777 cm–1. C21H30O2Si: calcd. C 73.63,
H 8.20; found C 73.34, H 7.95. Derivatization of alcohol with (R)-
(–)-2-methoxy-2-phenylacetic acid: 1H NMR (400 MHz, CDCl3): δ
= 4.29 (m, minor), 4.16 (dt, J = 7.5, 9.6 Hz, major), 3.22 (s, 81%),
3.16 (s, 19%) ppm. Derivatization of alcohol with (S)-(+)-2-meth-
oxy-2-phenylacetic acid: 1H NMR (400 MHz, CDCl3): δ = 4.39 (m,
major), 4.25 (dt, J = 7.5, 9.6 Hz, minor), 3.31 (s, 20%), 3.25 (s,
80%) ppm.
(1R,2S)-trans-2-(9-Anthracenyl)-1-cyclohexanol (19): The product
was isolated by FC (pentane/diethyl ether, 80:20) as a white solid.
1H NMR (400 MHz, CDCl3): δ = 8.61–7.54 (m, 8 H), 4.02 (dt, J
= 4.4, 10.1 Hz, 1 H), 4.16 (ddd, J = 4, 10.2, 12.6 Hz, 1 H), 2.57
(m, 1 H), 2.35 (m, 1 H), 2.03 (m, 3 H), 1.67 (m, 4 H) ppm. 13C
NMR (200 MHz, CDCl3): δ = 134.5, 132.4, 130.2, 129.7, 126.5,
125.0, 124.9, 126.2, 72.8, 48.7, 32.0, 30.1, 27.3, 25.5 ppm. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, CDCl3]: δ = 143.35 (s, 70.5%), 139.76 (s, 29.5%) ppm. ee =
41%. [α]2D0 = –10.4 (c = 1.48, MeOH) for 41% ee. M.p. 69–70 °C.
(1R,2S)-trans-2-Phenyl-1-cycloheptanol (30): The product was iso-
lated by FC (pentane/diethyl ether, 80:20) as acolorless oil. 1H
NMR (400 MHz, CDCl3): δ = 7.37–7.25 (m, 5 H), 3.79 (dt, J =
3.5, 13.7 Hz, 1 H), 2.60 (dt, J = 3.1, 9.7 Hz, 1 H), 2.03–1.69 (2 m,
11 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 145.7, 128.7, 127.6,
126.5, 77.7 (C-1), 55.3 (C-2), 35.2, 32.0, 27.3, 26.7, 21.8. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, CDCl3]: δ = 143.08 (s, 64.5%), 141.20 (s, 35.5%) ppm; ee =
IR: ν = 3483, 2929, 1623, 1449, 1028, 741 cm–1.
˜
(1R,2S)-trans-2-Phenyl-1-cyclopentanol (26):[56] The product was
1
isolated by FC (pentane/diethyl ether, 80:20) as a colorless oil. H
NMR (400 MHz, CDCl3): δ = 7.32–7.26 (m, 4 H), 4.16 (q, J =
7.24 Hz, 1 H), 2.89 (dt, J = 7.5, 9.5 Hz, 1 H), 2.14 (m, 2 H), 1.80
(m, 6 H) ppm. 13C NMR (200 MHz, CDCl3): δ = 141.9, 127.1,
126.0, 124.9, 78.9, 52.9, 32.4, 30.4, 20.3 ppm. 31P NMR [162MHz,
(1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine,
CDCl3]: δ = 141.85 (s, 68.5%), 138.69 (s, 31.5%) ppm. ee = 37%.
[α]2D0 = –29.8 (c = 1.34, EtOH) for 37% ee {ref.[57] [α]2D0 = +83.6 (c
= 1.565, EtOH) for (1S, 2R)-28, 99% ee}
29%. [α]2D0 = +2.52 (c = 0.99, CHCl ) for 29% ee. IR: ν = 3400,
˜
3
2910, 1440, 1010, 750, 690 cm–1. C13H18O: calcd. C 82.06, H 8.41;
found C 82.18, H 8.24. Derivatization of alcohol with (R)-(–)-2-
1
methoxy-2-phenylacetic acid: H NMR (400 MHz, CDCl3): δ =
3.26 (s, 65%), 3.11 (s, 35%), 2.85 (dt, J = 3.1, 10 Hz, minor), 2.80
(dt, J = 3.1, 10 Hz, major) ppm. Derivatization of alcohol with (S)-
(+)-2-methoxy-2-phenylacetic acid: 1H NMR (400 MHz, CDCl3):
δ = 3.16 (s, 34%), 3.11 (s, 66%), 2.85 (dt, J = 3.1, 10 Hz, major),
2.79 (dt, J = 3.1, 10 Hz, minor) ppm.
(1R,2S)-trans-2-(1-Naphthyl)-1-cyclopentanol (27): The product was
isolated by FC (pentane/diethyl ether, 80:20) as a white solid. 1H
NMR (400 MHz, CDCl3): δ = 8.25–7.53 (m, 7 H), 3.62 (q, J =
6.4 Hz, 1 H), 3.82 (dt, J = 6.7, 8.4 Hz, 1 H), 2.39–1.89 (m, 7 H)
ppm. 13C NMR (100 MHz, CDCl3): δ = 143.3, 139.9, 132.9, 129.2,
126.0, 125.9, 124.2, 122.8, 79.7, 49.5, 34.4, 32.1, 22.4 ppm. 31P
NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-eth-
anediamine, CDCl3]: δ = 141.90 (s, 76.5%), 138.60 (s, 23.5%) ppm.
ee = 53%. [α]2D0 = –18.7 (c = 1.05, CHCl3) for 53% ee. M.p. 71–
(1R,2S)-trans-2-(1-Naphthyl)-1-cycloheptanol (31): The product was
isolated by FC (pentane/diethyl ether, 80:20) as a white solid. 1H
NMR (400 MHz, CDCl3): δ = 8.22–7.53 (m, 7 H), 4.16 (br. s, 1
H), 3.58 (br. s, 1 H), 2.15–1.75 (m, 11 H) ppm. 13C NMR
(100 MHz, CDCl3): δ = 142.4, 132.5, 129.3, 127.3, 126.4, 126.1,
126.0, 123.8, 77.2, 48.7, 36.0, 33.1, 28.0, 27.4, 22.4 ppm. 31P NMR
[162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanedi-
amine, in the presence of S8, CDCl3]: δ = 143.08 (s, 25.5%), 81.16
72 °C. IR: ν = 3244, 2952, 1596, 1397, 1056, 778 cm–1. C H O:
˜
15 16
calcd. C 84.87, H 7.60; found C 84.98, H 7.53.
(1R,2S,5S)-trans-4-(tert-Butyldimethylsiloxy)-2-(1-naphthyl)-1-cy- (s, 74.5%) ppm; ee = 49%. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-
clopentanol (28): The product was isolated by FC (pentane/diethyl
ether, 80:20) as a colorless oil. H NMR (400 MHz, CDCl3): δ =
dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 140.57 (br.
1
s, minor), 143.34 (br. s, major) ppm. [α]2D0 = –6.09 (c = 1.05, CHCl3)
8.29–7.50 (m, 7 H), 4.68 (q, J = 7.2 Hz, 1 H), 4.57 (m, 1 H), 3.76
(dt, J = 7.5, 9.0 Hz, 1 H), 2.58 (m, 1 H), 2.08 (m, 1 H), 1.94 (m, 1
for 49 % ee. M.p. 98–99 °C. IR: ν = 3271, 2928, 1596, 1442, 1062,
˜
777 cm–1. C17H20O: calcd. C 84.96, H 8.39; found C 84.82, H 8.14.
Eur. J. Org. Chem. 2005, 1354–1366
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1363

E. Vrancken, A. Alexakis, P. Mangeney
FULL PAPER
(1R,2S)-trans-2-Phenyl-1-cyclooctanol (32): The product was iso-
formed according to the general procedure A but using (4S)-9 in-
stead of (–)-sparteine. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-di-
lated by FC (pentane/diethyl ether, 80:20) as a colorless oil. 1H
NMR (400 MHz, CDCl3): δ = 7.28 (m, 5 H), 3.95 (m, 1 H), 2.76 methyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 143.02 (s,
(dt, J = 5.6, 10 Hz, 1 H), 2.02–1.68 (m, 13 H) ppm. 13C NMR
(100 MHz, CDCl3): δ = 146.2, 129.9, 129.3, 128.4, 127.0, 76.4, 51.7,
23%), 141.71 (s, 77%) ppm; ee = 54%. [α]2D0 = +26.4 (c = 1.26,
benzene) for 54% ee {ref.[5] [α]2D0 = –22.4 (c = 1.26, benzene) for
32.9, 32.4, 27.5, 27.4, 25.7, 23.5 ppm. 31P NMR [162MHz, (1R,2R)- 47% ee}.
(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ =
1-Cycloocta-2,5-dienol (36):[58] A solution of sBuLi (3 mL, 1.3 m in
143.79 (s, 81%), 142.16 (s, 19%) ppm. ee = 62%.[α]2D0 = +4.95 (c =
hexane, 4 mmol) was added dropwise to a cooled (–90 °C) solution
of (–)-sparteine (0.92 mL, 4 mmol) in Et2O (10 mL) under argon.
After stirring for 30 min at this temperature, a solution of 1,5-cy-
clooctadiene oxide 35 (248 mg, 2 mmol) in Et2O (2 mL) was slowly
added dropwise (over 5 min). After stirring at –90 °C for 6 h, the
reaction was quenched with MeOH (2 mL) and an aqueous solu-
tion of 5% H2SO4 (10 mL). The aqueous layer was extracted with
Et2O (3×50 mL), the combined organic phases were washed with
brine, dried with MgSO4, and filtered off. The solvents were re-
moved under reduced pressure and the product was purified by
column chromatography on silica gel with pentane/diethyl ether
(75:25) as eluent to afford alcohol 36 (186 mg, 75%) as a colorless
oil. 1H NMR (400 MHz, CDCl3): δ = 5.65 (m, 2 H), 5.51 (m, 1 H),
5.37 (m, 1 H), 4.91 (m, 1 H), 2.84 (br. s, 2 H), 2.55 (m, 1 H), 2.09
(m, 1 H), 1.86 (m, 2 H), 1.42 (m, 1 H) ppm. 13C NMR (100 MHz,
CDCl3): δ = 134.0, 129.3, 128.9, 127.4, 69.3, 31.8, 29.3, 23.4 ppm.
31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-
ethanediamine, in the presence of S8, CDCl3]: δ = 83.20 (s, 81%),
83.03 (s, 19%) ppm. ee = 62%. [α]2D0 = –45.95 (c = 1.48, CHCl3)
for 62 % ee. C8H12O: calcd. C 77.38, H 9.74; found C 77.30, H 9.90.
1.07, CHCl ) for 62% ee. IR: ν = 3960, 2930, 1640, 1470, 1052,
˜
3
740 cm–1. C14H20O: calcd. C 82.30, H 9.87; found C 82.12, H 9.75.
Derivatization of alcohol with (R)-(–)-2-methoxy-2-phenylacetic
acid: 1H NMR (400 MHz, CDCl3): δ = 3.10 (s, 78%), 3.08 (s, 22%),
3.05 (ddd, minor), 3.02 (ddd, J = 2.6, 7.8, 11 Hz, major) ppm.
Derivatization of alcohol with (S)-(+)-2-methoxy-2-phenylacetic
acid: 1H NMR (400 MHz, CDCl3): δ = 3.10 (s, 21%), 3.08 (s, 79%),
3.04 (ddd, J = 2.6, 7.8, 11 Hz, major), 3.02 (ddd, minor) ppm.
(1R,2S)-trans-2-(1-Naphthyl)norbornanol (33): The product was iso-
lated by FC (pentane/diethyl ether, 80:20) as aviscous oil. 1H NMR
(400 MHz, CDCl3): δ = 8.12–7.51 (m, 7 H), 4.07 (br. s, 1 H), 3.58
(dd, J = 6.8, 8.8 Hz, 1 H), 2.54 (br. s, 1 H), 2.20–1.49 (m, 7 H)
ppm. 13C NMR (200 MHz, CDCl3): δ = 142.1, 134.6, 132.7, 129.4,
127.0, 126.2, 125.8, 124.9, 123.5, 81.0, 45.6, 43.7, 41.9, 36.3, 28.8,
26.0 ppm. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-di-
phenyl-1,2-ethanediamine, CDCl3]: δ = 140.42 (s, 33%), 141.12 (s,
67%) ppm. ee = 34%. [α]2D0 = –14.07 (c = 1.6, CHCl3) for 34% ee.
IR: ν = 3380, 2970, 1590, 1390, 1062, 770, 720 cm–1. C H O:
˜
17 18
calcd. C 85.67, H 7.61; found C 85.63, H 7.59. Derivatization of
alcohol with (R)-(–)-2-methoxy-2-phenylacetic acid: 1H NMR
(400 MHz, CDCl3): δ = 3.49 (t, J = 7.4 Hz, 36%), 3.44 (t, J =
7.4 Hz, 64%), 3.18 (s, 35.5%), 3.15 (s, 64.5%) ppm. Derivatization
of alcohol with (S)-(+)-2-methoxy-2-phenylacetic acid: 1H NMR
(400 MHz, CDCl3) 3.49 (t, J = 7.4 Hz, 62.5%), 3.44 (t, J = 7.4 Hz,
37.5%), 3.18 (s, 71%), 3.15 (s, 19%) ppm.
General Procedure B for the Carbenoid Rearrangement of meso-Ep-
oxidesby sBuLi/(–)-sparteine/BF3·Et2O: A solution of sBuLi (3 mL,
1.3 m in hexane, 4 mmol) was added dropwise, under argon, to a
cooled (–90 °C) solution of (–)-sparteine (0.92 mL, 4 mmol) in
Et2O (12 mL). After stirring for 30 min at this temperature, a solu-
tion of epoxide (2 mmol) in Et2O (2 mL) was injected with a sy-
ringe. A solution of BF3·Et2O (0.38 mL, 3 mmol) in Et2O (2 mL)
was slowly (over 5 min) added in order to maintain the temperature
at –90 °C. After stirring for 10 min, the reaction was quenched with
MeOH (2 mL) and Et3N (3 mL) and then the mixture was allowed
to reach room temperature. An aqueous solution of 5% H2SO4
(10 mL) was then slowly added to the reaction mixture. The aque-
ous layer was extracted with Et2O (3×50 mL), the combined or-
ganic phases were washed with brine, dried with MgSO4, and fil-
tered. The solvents were removed under reduced pressure and the
product was then subjected to FC on SiO2.
(1R,2S)-trans-2-Phenyl-5,6-cycloocten-1-ol (39): The product was
1
isolated by FC (pentane/diethyl ether, 80:20) as a colorless oil. H
NMR (400 MHz, CDCl3): δ = 7.15 (m, 5 H), 5.79 (ddd, J = 1, 5.4,
16.8 Hz, 1 H), 5.60 (m, 1 H), 4.08 (ddd, J = 3.6, 5.7, 9.3 Hz, 1 H),
2.93 (ddd, J = 3.9, 9.4, 11.2 Hz), 1.67–2.34 (m, 9 H) ppm. 13C
NMR (200 MHz, CDCl3): δ = 144.4, 131.9, 129.2, 128.8, 127.2,
74.6, 49.8, 34.1, 34.0, 25.2, 24.9 ppm. 31P NMR [162MHz, (1R,2R)-
(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ =
145.59 (s, 75.5%), 142.35 (s, 24.5%) ppm; ee = 51%. [α]2D0 = +14.85
(c = 1.38, CHCl ) for 51% ee. IR: ν = 3450, 2980, 1610, 1410, 1060,
˜
3
720 cm–1. C14H28O: calcd. C 83.12, H 8.97; found C 81.48, H 10.70.
(1S)-endo-cis-Bicyclo[3.3.0]octan-2-ol (34):[36] The reaction was per-
formed according to general procedure B. The product was isolated
by FC (pentane/diethyl ether, 80:20) as a colorless oil. The NMR
Nucleophilic Opening of Cyclohexene Oxide by nBuLi/8a/BF3·Et2O
To Give trans-2-Butyl-1-cyclohexanol (3): The reactions were per-
formed according to the general procedure A but using (1R,2S)-8a
instead of (–)-sparteine. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-
dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 143.11 (s,
71%), 139.78 (s, 31%) ppm. ee = 38%. [α]2D0 = –22.5 (c = 1.46,
CHCl3) for 35% ee {ref.[5] [α]2D0 = –6 (c = 1.08, CHCl3) for 12%
ee}.
spectra of (1S)-34 are identical to those reported previously.[36] 31
P
NMR [162 MHz, (1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-eth-
anediamine, CDCl3]: δ = 138.72 (s, 19.5%), 140.76 (s, 80.5%) ppm.
ee = 61%. [α]2D0 = –12.52 (c = 1.025, CHCl3) for 61% ee {ref.[36]
[α]2D0 = –19 (c = 1.0, CHCl3) for 84% ee}.
General Procedure C for the Carbenoid Rearrangement of meso-Ep-
oxidesby Organolithium/(–)-Sparteine/BF3·Et2O: A solution of
sBuLi (3 mL, 1.3 m in hexane, 4 mmol) was added dropwise, under
argon, to a cooled (–90 °C) solution of (–)-sparteine (0.92 mL,
4 mmol) in Et2O (10 mL). After stirring at this temperature for
30 min, a solution of BF3·Et2O (0.38 mL, 3 mmol) in Et2O (2 mL)
was slowly (over 20 min) added in order to maintain the tempera-
ture at –90 °C. A solution of epoxide (2 mmol) in Et2O (2 mL) was
then slowly added dropwise (over 20 min). After stirring for 10 min,
the reaction was quenched with MeOH (2 mL) and an aqueous
solution of 5% H2SO4 (10 mL). The aqueous layer was extracted
Nucleophilic Opening of Cyclohexene Oxide by nBuLi/8b/BF3·Et2O
To Give trans-2-Butyl-1-cyclohexanol (3): The reactions were per-
formed according to the general procedure A but using (1S,2S)-8b
instead of (–)-sparteine. 31P NMR [162 MHz, (1R,2R)-(+)-N,NЈ-
dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ = 143.22 (s,
39%), 139.96 (s, 61%) ppm; ee = 22%. [α]2D0 = +12.6 (c = 1.48,
CHCl3) for 22% ee {ref.[5] [α]2D0 = –6 (c = 1.08, CHCl3) for 12%
ee}.
Nucleophilic Opening of Cyclohexene Oxide by PhLi/9/BF3·Et2O To
Give trans-2-Butyl-1-cyclohexanol (3): The reactions were per-
1364
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1354–1366

Enantioselective Desymmetrization of meso-Epoxides
FULL PAPER
[11]
[12]
[13]
[14]
[15]
H. C. Brown, U. S. Racherla, S. M. Singh, Tetrahedron Lett.
1984, 25, 2411–2414.
M. J. Eis, J. E. Wrobel, B. Ganem, J. Am. Chem. Soc. 1984,
106, 3693–3694.
A. Ghribi, A. Alexakis, J. F. Normant, Tetrahedron Lett. 1984,
25, 3075–3078.
B. H. Lipschutz, D. A. Parker, J. A. Koslowski, S. L. Ngyuen,
Tetrahedron Lett. 1984, 25, 5959–5962.
For a general overview see: D. M. Hodgson, M. A. H. Stent,
Topics in Organometallic Chemistry: Organolithiums in Enanti-
oselective Synthesis (Ed.: D. M. Hodgson), Springer, Heidel-
berg, 2003, vol 5, pp. 1–21.
with Et2O (3×50 mL), the combined organic phases were washed
with brine, dried with MgSO4, and filtered. The solvents were re-
moved under reduced pressure and the product was then subjected
to FC on SiO2.
Tricyclo[2.2.1.0]heptan-3-ol (40):[36] The reaction was performed ac-
cording to general procedure B. The NMR spectra of 40 are iden-
tical to those reported previously.[36] 31P NMR [162 MHz, (1R,2R)-
(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine, CDCl3]: δ =
138.86 (s, 52%), 137.02 (s, 48%) ppm. ee = 4%.
Cycloocta-2,5-dien-1-ol (36), (1S)-Bicyclo[3.3.1]heptan-3-ol (37),
and trans-8-sec-Butylcyclooct-4-en-1-ol (38): The reaction was per-
formed according to general procedure B. The residue was purified
by column chromatography on silica gel with pentane/diethyl ether
(75:25) as eluent. First to elute was alcohol 36 (173 mg, 7%, Rf =
0.33), then 37 (1.6 g, 65%, Rf = 0.25) and 38 (291 mg, 8%, Rf =
0.42) as colorless oils. (1S)-37:[59] 1H NMR (400 MHz, CDCl3): δ
= 5.78 (m, 1 H), 5.45 (ddd, J = 2.7, 7.6, 11 Hz, 1 H), 4.28 (m, 1
H), 1.39–2.03 (m, 7 H), 0.87 (dt, J = 4, 8.9 Hz, 1 H), 0.51 (dt, J =
4, 6.1 Hz, 1 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 129.6,
127.9, 72.4, 34.2, 26.9, 24.3, 15.9, 11.8 ppm. 31P NMR [162 MHz,
(1R,2R)-(+)-N,NЈ-dimethyl-1,2-diphenyl-1,2-ethanediamine, in the
presence of S8, CDCl3]: δ = 82.52 (s, 87.5%), 82.11 (s, 12.5%) ppm.
ee = 75%. [α]2D0 = –86.16 (c = 1.03, CHCl3) for 75% ee. Derivati-
zation of alcohol with (R)-(–)-2-methoxy-2-phenylacetic acid: 1H
NMR (400 MHz, CDCl3): δ = 3.45 (s, 14%), 3.44 (s, 86%), 0.95
(dt, 86%), 0.73 (dt, 14%), 0.57 (dt, 87%), 0.43 (dt, 13%) ppm.
Derivatization of alcohol with (S)-(+)-2-methoxy-2-phenylacetic
acid: 1H NMR (400 MHz, CDCl3): δ = 3.45 (s, 84%), 3.43 (s, 16%),
0.95 (dt, 13%), 0.73 (dt, 87%), 0.57 (dt, 15%), 0.43 (dt, 85%) ppm.
38: 1H NMR (400 MHz, CDCl3): δ = 5.66 (m, 1 H), 5.52 (m, 1 H),
3.75 (m, 1 H), 4.28 (m, 1 H), 1.24–2.27 (m, 14 H), 0.89 (t, J =
7.2 Hz, 3 H), 0.82 (d, J = 6.7 Hz, 3 H) ppm. 13C NMR (100 MHz,
CDCl3): δ = 131.1, 130.5, 128.8, 128.2, 73.6, 47.2, 44.2, 36.6, 35.1,
34.3, 28.2, 25.6, 25.4, 25.0, 24.3, 24.1, 18.2, 14.2, 12.9, 12.7 ppm.
[16]
[17]
P. Müller, P. Nury, Org. Lett. 2000, 2, 2845–2847.
P. Müller, P. Nury, G. Bernardinelli, Eur. J. Org. Chem. 2001,
4137–4147.
For a review on (–)-sparteine, see: D. Hoppe, T. Hense, Angew.
Chem. Int. Ed. Engl. 1997, 36, 2282–2316.
The enantiomeric excess was measured according to: A.
Alexakis, S. Mutti, P. Mangeney, J. F. Normant, Tetrahedron:
Asymmetry 1990, 1, 437–440.
The absolute configuration was determined to be (1R,2S) by
polarimetry and comparison of the optical rotation with the
known value (see Exp. Sect.).
A. Thompson, E. G. Corley, M. F. Huntington, E. J. J. Gra-
bowski, J. F. Remenar, D. B. Collum, J. Am. Chem. Soc. 1998,
120, 2028–2038.
F. Xu, R. A. Reamer, R. Tillyer, J. M. Cummins, E. J. J. Gra-
bowski, P. J. Reider, D. B. Collum, J. C. Huffman, J. Am. Chem.
Soc. 2000, 122, 11 212–11 218.
T. F. Briggs, M. D. Winemiller, D. B. Collum, R. L. Par-
sons, Jr., A. H. Davulcu, G. D. Harris, J. M. Fortunak, P. N.
Confalone, J. Am. Chem. Soc. 2004, 126, 5427–5435.
J. K. Whitesell, Chem. Rev. 1992, 92, 953–964.
The absolute configuration was determined to be (1R,2S) by
polarimetry and comparison of the optical rotation with the
known value (see Exp. Sect.).
B. M. Trost, J. L. Belletire, S. Godleski, P. G. McDougal, J. M.
Balkovec, J. J. Baldwin, M. E. Christy, G. S. Ponticello, S. L.
Varga, J. P. Springer, J. Org. Chem. 1986, 51, 2370–2374.
A. Alexakis, E. Vrancken, P. Mangeney, J. Chem. Soc., Perkin
Trans. 1 2000, 2254–3355.
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
IR: ν = 3380, 2903, 1450, 1030, 740 cm–1. C H O: calcd. C 79.06,
˜
12 22
[27]
[28]
H 12.16; found C 79.08, H 12.14.
For a general overview of lithiated epoxide rearrangement, see:
D. M. Hodgson, E. Gras, K. Tomooka, Topics in Organometal-
lic Chemistry: Organolithiums in Enantioselective Synthesis
(Ed.: D. M. Hodgson), Springer, Heidelberg, 2003, vol. 5, pp.
217–250.
Acknowledgments
The authors thank the MESR for a grant.
[29]
[30]
A. C. Cope, H. H. Lee, H. E. Petree, J. Am. Chem. Soc. 1958,
80, 2849–2852.
J. K. Crandall, M. Apparu, Org. React.(N. Y.) 1983, 29, 345–
443.
[1] D. K. B. Aubrecht, M. D. Winemiller, D. B. Collum, J. Am.
Chem. Soc. 2000, 122, 11 084–11 089 and references cited
therein.
[2] For a comprehensive review see: M. C. Willis, J. Chem. Soc.,
Perkin Trans. 1 1999, 1765–1784.
[31]
[32]
[33]
T. Satoh, Chem. Rev. 1996, 96, 3303–3325.
K. M. Morgan, S. Gronert, J. Org. Chem. 2000, 65, 1461–1466.
A. Ramirez, D. B. Collum, J. Am. Chem. Soc. 1999, 121,
11 114–11 121.
D. M. Hodgson, G. P. Lee, Chem. Commun. 1996, 1015–1016.
D. M. Hodgson, G. P. Lee, Tetrahedron: Asymmetry 1997, 8,
2303–2306.
D. M. Hodgson, G. P. Lee, R. E. Marriott, A. J. Thompson, R.
Wisedale, J. Withrington, J. Chem. Soc., Perkin Trans. 1 1998,
2151–2161.
S. H. Wiedemann, A. Ramirez, D. B. Collum, J. Am. Chem.
Soc. 2003, 125, 1583–15901.
For a recent review of an asymmetric version, see: A. Magnus,
S. K. Bertilsson, P. G. Andersson, Chem. Soc. Rev. 2002, 31,
223–229 and references cited therein.
J. K. Crandall, L. H. Chang, J. Org. Chem. 1967, 32, 532–536.
J. K. Crandall, J. Org. Chem. 1964, 29, 2830–2833.
J. K. Crandall, L. C. Crawley, D. B. Banks, L. C. Lin, J. Org.
Chem. 1971, 36, 510–513.
[3] For a general overview of the reactivity of epoxides with
organolithium reagents see: F. Chemla, E. Vrancken, Patai
Series:The Chemistry of Organolithium Compounds (Ed.: Z.
Rappoport), Wiley, Chichester, 2004, pp. 1165–1242.
[4] M. Mizuno, M. Kanai, A. Iida, K. Tomioka, Tetrahedron:
Asymmetry 1996, 7, 2483–2484.
[34]
[35]
[36]
[5] M. Mizuno, M. Kanai, A. Iida, K. Tomioka, Tetrahedron 1997,
53, 10 699–10 708.
[37]
[38]
[6] A. Alexakis, E. Vrancken, P. Mangeney, Synlett 1998, 1165–
1166.
[7] A. Alexakis, A. Tomassini, S. Leconte, Tetrahedron 2004, 60,
9479–9484.
[8] N. Oguni, Y. Miyagi, K. Itoh, Tetrahedron Lett. 1998, 39, 9023–
9026.
[39]
[40]
[41]
[9] C. Zhu, M. Yang, J. Sun, Y. Zhu, Y. Pan, Synlett 2004, 465–
468.
[10] M. Yamaguchi, I. Hirao, Tetrahedron Lett. 1983, 24, 391–394.
Eur. J. Org. Chem. 2005, 1354–1366
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1365

E. Vrancken, A. Alexakis, P. Mangeney
FULL PAPER
[42] D. M. Hodgson, R. Wisedale, Tetrahedron: Asymmetry 1996,
7, 1275–1276.
[43] J. K. Crandall, D. B. Banks, R. A. Colyer, R. J. Watkins, J. P.
Arrington, J. Org. Chem. 1968, 33, 423–425.
[51] D. Basavaiah, P. D. Rao, Tetrahedron: Asymmetry 1994, 5, 223–
234.
[52] H. Timmerman, R. F. Rekker, W. Th. Nauta, Recl. Trav. Chim.
Pays-Bas 1965, 84, 1348–1356.
[44] M. Asami, Bull. Chem. Soc. Jpn. 1990, 63, 1402–1408.
[45] Prepared according to: A. I. Meyers, G. S. Poindexter, S. Brich,
J. Org. Chem. 1978, 43, 892–898.
[46] J. K. Whitesell, R. M. Lawrence, Chimia 1986, 40, 318–321.
[47] A Alexakis, D. Jachiet, J. F. Normant, Tetrahedron 1986, 42,
5607–5619.
[48] R. K. Sehgal, R. U. Koenigsberger, T. J. Howard, J. Chem.
Soc., Perkin Trans. 1 1976, 191–195.
[49] J. Barluenga, F. Alvarez, J. M. Concellon, P. Bernad, M. Yus,
Synthesis 1987, 318–320.
[53] B. C. McKusick, J. Am. Chem. Soc. 1948, 70, 1976–1977.
[54] F. Dumas, B. Mezrhab, J. d’Angelo, J. Org. Chem. 1996, 61,
2293–2304.
[55] J. W. Cook, C. L. Hewett, C. A. Lawrence, J. Chem. Soc. 1936,
71–80.
[56] L. A. Asuncion, J. E. Baldwin, J. Org. Chem. 1995, 60, 5778–
5784.
[57] H. C. Brown, J. V. N. Vara Prasad, A. K. Gupta, R. K. Bakshi,
J. Org. Chem. 1987, 52, 310–311.
[58] E. S. Koltun, S. R. Kass, Synthesis 2000, 1366–1368.
[59] M. Apparu, M. Barelle, Bull. Soc. Chim. Fr. 1977, 947–950.
Received October 18, 2004
[50] J. M. Barks, G. G. Weingarten, D. W. Knight, J. Chem. Soc.,
Perkin Trans. 1 2000, 3469–3476.
1366
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1354–1366